The structure of innate vocalizations in Foxp2-deficient mouse pups

Abstract

Heterozygous mutations of the human FOXP2 gene are implicated in a severe speech and language disorder. Aetiological mutations of murine Foxp2 yield abnormal synaptic plasticity and impaired motor-skill learning in mutant mice, while knockdown of the avian orthologue in songbirds interferes with auditory-guided vocal learning. Here, we investigate influences of two distinct Foxp2 point mutations on vocalizations of 4-day-old mouse pups (Mus musculus). The R552H missense mutation is identical to that causing speech and language deficits in a large well-studied human family, while the S321X nonsense mutation represents a null allele that does not produce Foxp2 protein. We ask whether vocalizations, based solely on innate mechanisms of production, are affected by these alternative Foxp2 mutations. Sound recordings were taken in two different situations: isolation and distress, eliciting a range of call types, including broadband vocalizations of varying noise content, ultrasonic whistles and clicks. Sound production rates and several acoustic parameters showed that, despite absence of functional Foxp2, homozygous mutants could vocalize all types of sounds in a normal temporal pattern, but only at comparably low intensities. We suggest that altered vocal output of these homozygotes may be secondary to developmental delays and somatic weakness. Heterozygous mutants did not differ from wild-types in any of the measures that we studied (R552H ) or in only a few (S321X ), which were in the range of differences routinely observed for different mouse strains. Thus, Foxp2 is not essential for the innate production of emotional vocalizations with largely normal acoustic properties by mouse pups.

Figure 1. Body weight of pups of both FOXP2 mouse lines (R552H and S321X)

Homozygous mutants of both lines weighed significantly less compared with their wild-type and heterozygous littermates. The number of animals per group is shown in the top row of the figure.

Figure 2. Example spectrograms of series of ultrasounds (USIs) and clicks from isolated pups

The genotype of the vocalizing pup is indicated below the spectrograms. (a, b) USIs, partly with clicks; (c) clicks without USIs. Some USIs (a, b, d, e) end with a frequency jump.

Figure 3. Example spectrograms of DCs alternating with USD and clicks

The genotype of the vocalizing pup is indicated below the spectrograms. The DCs may show a high-noise content (high-noise DC), a medium-noise content (medium-noise DC) or a low-noise content (low-noise DC). USDs may start with a click and may end with a frequency jump.

Figure 5. Percentages of the emitted DC types according to their noise content

In homozygotes of both mouse lines [(a)R552 H and (b)S321 X], DCs of high-, medium- or low-noise content occurred at similar rates of about 30%. Wild-type and heterozygous animals of both mouse lines produced more high-noise than medium-noise or low-noise DCs; the wild-types also more medium-noise than low-noise DCs. Sounds of 10–12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 4. Number of vocalizations emitted by S321X mutants

(a) USIs. Wild-types and heterozygotes produced similar USI rates during the 15 min of recording, while only 4 of the 12 tested homozygotes vocalized under these conditions, producing just a few USIs. (b) Number of DCs and USDs emitted in the distress situation. The rates of DCs did not depend on the genetic background. The ratio of USDs to DCs varied among the different groups, with lowest numbers of USDs emitted by homozygotes. Data for R552H lines are shown in Fig. S1. Sounds of 11 or 12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 6. Ultrasounds (USDs) and clicks produced by R552H and S321X mutants in the distress condition

(a) About 80% of the USDs from the wild-type and heterozygous animals of both mouse lines contained clicks, while only about 50% of the USDs from the respective homozygotes had clicks. Percentage of USIs with clicks is shown in Fig. S3. (b) Wild-type and heterozygous pups emitted more USDs than clicks, while this was reversed in homozygotes. The sum of the emitted clicks and USDs, however, did not differ between genotypes. Data for R552H lines are shown in Fig. S3. Sounds of 8 to 12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 7. Durations of DCs and of USDs or USIs

(a) The duration of DCs was similar for the three groups of the R552H mouse line, while S321X homozygotes had shorter DCs than heterozygotes (b). USDs of homozygotes of both mouse lines were shorter than those of wild-types and heterozygotes of the respective lines (a, b). There were no differences in duration between USIs of wild-type and heterozygotes of both mouse lines (c). Sounds of 8–12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 8. Duration of inter-call intervals in series of DCs and series of USIs

(a, b) Inter-call intervals in series of DCs did not differ among the genotypes of either line. (c) Inter-call intervals in series of USIs of the S321X heterozygotes were longer than those of their wild-type littermates. USI inter-call intervals did not differ between R552H heterozygotes and their wild-type littermates. Mostly, sounds of 12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 9. Percentages of ultrasounds with frequency jumps in isolation [USIs, (a)] or in distress [USDs, (b)]

There were no differences between the genotypes of the R552H mouse line with regard to the percentages of frequency jumps in USIs or USDs. S321X wild-types, however, emitted more USIs and USDs with frequency jumps than other genotypes. Sounds of 8–12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 10. Peak SPL of DCs and USDs

(a, b) DCs and USDs of homozygotes of both mouse lines were softer than the calls emitted by the respective heterozygous and wild-type animals. Sounds of 8–12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 11. Peak SPL of DCs in relation to their noise content

(a, b) In wild-types of both mouse lines and R552H heterozygotes, high-noise DCs were louder than the low-noise DCs. Homozygous mutants of both mouse lines produced the softest calls without significant differences in SPL between the DCs of different noise content. Sounds of 6–12 animals per group could be analysed (see numbers in the top row of the figure).

Figure 12. Spearman correlation analysis between body weight and number of recorded USIs of the pups

The correlation coefficients (r) and corresponding P values are indicated. The regression lines (y = y₀ + a ^*x) with y₀ = −521.9, a = 290.0 for the R552H pups (a) and y₀ = −227.8, a = 158.2 for the S321X pups (b) describe these correlations. On average, the number of recorded USIs increases with increasing body weight for both mouse lines.